PET 3-DIMENSIONAL DATA QUANTIFICATION

本研究的主要目标是量化PET数据

The primary goal of the proposed research is to quantitate PET data acquired in 3-D-mode. The outstanding feature of PET is that it enables researchers and clinicians to assign a quantitative value or rate to an in vivo physiological process. These processes include such things as glucose utilization by tumors, blood supply to the heart, or-the ability of the brain to use neurotransmitters. By removing the collimating septa and acquiring PET data in 3D-mode, the sensitivity is increased by a factor of -6, compared to the conventional slice-mode. If 3D PET could be used on a quantitative basis, the 6-fold increase in sensitivity would translate directly to either a 6-fold decrease in patient dose, a 6-fold increase in statistics, or a 6-fold increase in patient throughput. While data can be acquired in 3D-mode today, it is not yet possible to assign a quantitative value to this data. The difficulties in quantitating 3D PET stem from the lack of collimation, which permits a greater number of undesirable scattered gama-rays to be detected. A tradeoff is made: more "true" events are detected, but more scattered events are also detected, which makes quantitation more difficult. The research proposed here seeks to develop methods and algorithms to enable 3D PET quantitation. The specific areas to be addressed include: detector calibration (assigning an accurate muCi/ml tissue to each image pixel); detector normalization (correcting for differing sensitivities of individual detectors); scatter correction: (correcting for the undesirable scattered events); attenuation correction (since radioactive events originating deep in a tissue mass will be scattered more than those from a superficial location); and image reconstruction (transforming sinogram data into a recognizable image).